Imagine that you’re a patient being treated for prostate cancer—or that you are a physician treating such a patient—and the treatment is not going well. Seeking more effective care, you’ve signed up for alerts on new treatments and you receive this press release (only snippets quoted):
Foundation Medicine, Inc. today announced that it has received approval from the U.S. Food and Drug Administration (FDA) for FoundationOne®CDx to be used as a companion diagnostic for LYNPARZA® (olaparib), which was also approved today in the U.S. for adult patients with deleterious or suspected deleterious germline or somatic homologous recombination repair (HRR) gene-mutated metastatic castration-resistant prostate cancer (mCRPC) who have progressed following prior treatment with enzalutamide or abiraterone….“This therapy and companion diagnostic approval underscores the value of comprehensive genomic profiling in advanced cancer patients as it validates our ability to identify alterations in the 14 HRR pathway genes within FoundationOne CDx’s 324 gene panel that indicate a patient may be eligible for treatment with Lynparza, a process not possible through single gene or hot spot testing,” said Brian Alexander, M.D., M.P.H., chief medical officer at Foundation Medicine.
Imagine trying to make sense of all these unfamiliar terms. Imagine that you had to decide—ideally with the help of a knowledgeable physician—whether to take the diagnostic test and the related drug.
Bryce Olson, Heat-Seeking Prostate Cancer Missile
Bryce Olson was in exactly this position. He’s had prostate cancer for over seven years, and the course of the disease has not run smoothly for him. As an article in Wired explains, when the cancer was diagnosed, the “standard of care” was four months of toxic chemotherapy treatment. Olson underwent the chemo with all the usual horrible side effects, but his cancer still spread to his bones. He read that the median five-year survival rate for his cancer and stage was only 28%.
So Olson, then an Intel marketing executive, morphed into a “Citizen Oncologist.” He devoured and made sense of cancer research findings, and learned several important facts about prostate cancer:
- All prostate cancers are not alike, and his own type wouldn’t have as durable a response to either the chemotherapy or another “standard of care” treatment—drugs to neutralize the effects of testosterone on the body;
- Prostate cancers sometimes involve genetic mutations that can only be discovered through genomic screening of a patient’s tumors or blood;
- Cancer research—prostate and other types—is progressing very rapidly in research labs, but only slowly in clinical settings.
Since Olson was rapidly becoming an expert in cancer treatment, he asked to be moved into Intel’s Health and Life Sciences Technology group, where he is now a Global Strategist. Perhaps because of his Intel affiliation, he was able to get his tumor tissue DNA sequenced in January 2015 at Oregon Health & Science University medical center, which had one of the earliest capabilities for such sequencing. However, initial doubt was cast on the value of the sequencing as there were no drugs approved by the FDA for any mutations in his cancer that the sequencing might find.
From the tumor sequencing, Olson learned that he had an abnormal PI3K signaling pathway, which helps to regulate cell growth in cancer. He also had a mutation in the PTEN tumor suppressor gene; that also caused his cancer to grow unusually quickly. His Oregon doctor helped him find a clinical trial for a drug known as a PI3K inhibitor, developed by Verastem Oncology. The trial was based at Cedars Sinai hospital in Los Angeles, but Olson was willing to make the trip from Portland every few weeks. The drug seemed to work, and Olson’s cancer went into remission for a couple of years. But the trial ended because the drug was deemed toxic to some patients’ hearts. At the same time the drug supply ran out for Olson, his cancer came back. He got into another trial—this one in Oklahoma City—for a different P13K inhibitor, but it didn’t seem to work on his tumor. It was spreading to other bones in his body. Olson had learned firsthand that some cancer tumors mutate rapidly, and the effects of any targeted treatment strategy can be short-lived.
Olson wasn’t discouraged enough to stop looking for treatments. He knew there were many researchers testing many drugs that—alone or in combination—might stop the growth of his cancer. He was seeking an alternative to the standard of care, and he began to look for an oncologist whose data-driven approach matched his own. It wasn’t easy; as the Wired article put it,
As with any new technology, there are roadblocks to adoption: Clinicians need to master a steep learning curve that includes keeping track of new tests, drug approvals, clinical trials, and constantly changing treatment guidelines.
Dr. Rana McKay at UCSD
Bryce Olson found his “dream oncologist” in Dr. Rana McKay at the University of California San Diego (UCSD) Health. She is a research oncologist who conducts clinical trials (eight of them in mid-2021), and she focuses particularly on the drivers of prostate cancer. She’s a big fan of data-driven precision medicine. And Olson found her warm and dedicated to his cause; she spent an hour and a half with him in their first meeting in her clinic. “I will fly to San Diego for the rest of my life to have someone who will give me that much time,” Olson told me. “We’re always exploring new treatment ideas. She is so intellectually curious about what’s driving my cancer. And she is incredibly well-connected to other researchers in this area. Olson says that McKay balances out some of his aggressiveness: “I am more gonzo about taking drugs that have never been tested in humans. Rana is very focused on safety, which is good.” In short, Olson seems himself as the ideal of the data-driven patient, and says that McKay is “well-primed for the future patient who will be more like me.” He hopes she is learning something from him too, but he knows he is learning a lot from her.
Dr. McKay told me in an interview that oncology needs a new paradigm of how drugs enter clinical practice:
“It’s a pretty clunky, long, laborious process testing one drug at a time. We need to be more efficient at getting drugs to enter the clinic. The standard of care has its limitations for prostate cancer. Men run out of options, although they may still feel well. It’s hard when you’ve exhausted all the standard of care options, and there is nothing else on the shelf to try. With patients like Bryce we’re pushing the limits earlier. Ultimately, we hope to cure more patients and prevent the progression of their cancer.”
McKay described the approach that Bryce Olson has explored for several years now:
“We’re using a precision-based trial approach in prostate cancer. We take patients who have had sequencing, and based on select biomarkers we try to match them with a therapy. We want to learn what are the genomic drivers of disease—what’s really making this tumor grow? But the frustration in precision medicine is that we’ve adopted a strategy of targeting one or two mutations at any given time. However, tumors are heterogeneous. Additionally, some mutations don’t yet have effective drugs for targeting.”
In addition to the multiple drivers issue, McKay said that genetic testing for prostate cancers isn’t done enough, particularly outside of academic medical centers like hers. She also feels that the amount of information can be overwhelming for the general practitioner or the average oncologist:
“Advances in technology have led to overwhelming amount of data on tumor biology, patterns of disease, and outcomes for patients. How do you take all that information and distill it down into a therapy choice for each patient? At an academic cancer center like UCSD with robust precision medicine program and institutional Molecular Tumor Board there are tremendous resources available in helping better interpret clinical-grade sequencing data, but such resources may not be readily available across all oncology practices.”
Foundation Medicine—Genomics and AI vs. Cancer
There are several companies now that provide genomic sequencing of cancer tissue samples for biomarkers that may aid treatment. One of the most advanced is Foundation Medicine Inc. (FMI), based in Cambridge, MA (and owned by the pharmaceutical giant Roche). FMI is using its vast genomic database and partnering with biopharma organizations to try to translate findings about genes and mutations into treatment options and clinical trial possibilities for cancer patients. Bryce Olson has had his cancer cells sequenced at several companies, including FMI.
FMI tries to match patients like Olson with treatment options and trials. As described in the press release above, it partners with pharma firms that have developed cancer treatment drugs to validate the tests’ ability to detect specific mutations and match appropriate patients to approved targeted therapies. It generates a report curated by genomics experts and pathologists for patients’ oncologists about the genomic mutations present in the patient’s tumor or blood (for example, BRCA2, commonly found in breast cancer, but also sometimes present in prostate cancer), the FDA-approved therapeutic options (for example, olaparib and rucaparib, both recommended for treatment of cancers with BRCA2 mutations), and other genomic biomarker findings of possible clinical relevance. The report also lists clinical trials a patient may be eligible for in which the treatment options are being tested.
FMI is gathering aggregations of de-identified patient genomic data from its own tests, and it plans to use that data at some point to train machine learning models to recommend treatments and trials. But Priti Hegde, FMI’s Chief Scientific Officer, told me that there is more human learning than machine learning at work today:
“We are gathering data on the response of patients to particular cancer treatments. But that takes years to play out. Typically, one needs millions of cases to train machine learning models, but so far we have only thousands or in some cases hundreds of thousands. We do use systems to recommend treatments and trials, but they are a rules-based approach that represents the knowledge generated from our large database and from the scientists at FMI.”
Dr. McKay said that the companies try to make their reports provider-friendly, but they are not specific to particular types of cancer, and they do not account for the patient’s personal cancer history, including prior treatments. She notes that the genetic screening companies, including FMI, are making progress at explaining treatment options to oncologists. It appears, however, that the physician would still need to know a good deal about the genomics of cancer in order to use the information effectively in patient care.
Moving Forward in Precision Cancer Treatment
Both Bryce Olson and Dr. McKay also are attempting to influence how others with cancer might have better outcomes. Dr. McKay would like to see not only new drugs and data, but also new approaches to providing expert guidance to cancer patients and their physicians. “Tumor boards” exist at many healthcare provider institutions; they are meetings where physicians of cancer patients discuss specific patients’ cases treatment plans. UCSD Health has established a “Molecular Tumor Board” of oncologists and genomic experts at the medical center to discuss cancer patients whose tumors have been analyzed with advanced genomic diagnostic tests.
McKay believes this is a good idea that could be expanded. She envisions a broad molecular tumor board with perhaps a hundred experts or more. Caris Life Sciences, another precision cancer care firm that does genomic sequencing, has such a board the provides the opportunity for oncologists to virtually interact with cancer experts to review molecular information on specific patient cases. FMI also conducts molecular tumor boards. McKay believes that companies could establish larger boards to complement the AI-based data interpretation that already exists. “A source of doctor-to-doctor contact on a patient’s situation could be very helpful.” She also feels that virtual platforms including AI capabilities could do much more in the future.
Olson’s cancer has been discussed at several tumor boards. But the one he found most helpful was a virtual discussion involving himself, Dr. McKay, three other oncologists from different medical centers, scientists focused on clinical development for prostate cancer at FMI, and a representative from Xcures, a startup that is using AI to match cancer patients with treatment options beyond the standard of care. The team discussed and debated 5 therapeutic approaches and got consensus on the top 3 in order.
As Bryce Olson continues to “hack this thing,” as he describes his personal fight with cancer, he’s looking for new approaches. He says that the information he received from his tumor sequencing was only the beginning; “I’ve gone well beyond that.” He’s focused on some new drugs coming out that are based on proteomics—identifying protein “hot spots” in cancer cells and finding drugs—often in combination—to degrade them. His concern is that his cancer is growing faster than the science is.
Olson is working with Intel to advance the state of the art in technology for cancer care. He’s also built relationships with influential cancer researchers and with CEOs of cancer startups like FMI and Xcures. He speaks publicly often about his own experience with cancer, and encourages patients and oncologists to exploit the precision treatment approaches based on genomics. He’s engaged in a war against the conservative “standard of care” that so many cancer patients receive. He commented:
I believe that everyone has to own their own healthcare. When I was following my doctor’s standard of care orders, I was not engaged, and was kind of depressed. When I started to individualize my care and work with personalized diagnostics that are unique to me, I became engaged, and got my swagger back. Being on the forefront of science and technology is where every patient should be.”